How to Track AI Traffic in GA4 with the New AI Assistant Channel
Find AI referrals, build useful reports, and understand what the numbers can tell you.
Google Analytics 4 now has a built-in way to group website visits from recognized AI assistants. To track AI traffic in GA4, open Traffic acquisition, choose Session default channel group, and look for AI Assistant. That gives you a starting point for seeing whether AI tools are sending people to your website.
Google announced the AI Assistant channel back on May 13, 2026, but most people never saw it. If you own a business or lead its marketing, the useful next step is understanding what those visitors do when they arrive.
We set up dedicated AI traffic reports for ClickIt CMO, and I thought others might want to know how we did it. Here is how the feature works, what deserves a closer look, and how we can help if GA4 setup is outside your comfort zone or your team’s capacity.
What changed in Google Analytics
GA4’s AI Assistant channel automatically groups visits when their referrer matches an AI assistant Google recognizes. Google’s announcement names ChatGPT, Gemini, and Claude among the examples. A referrer is information that can identify the site sending a visitor to yours.
For recognized AI referrals, GA4 assigns the medium ai-assistant and the channel AI Assistant. You can use that native classification without building your own list of AI domains first.
Your website still needs functioning Analytics tracking. The new channel organizes the source information GA4 receives; it does not fix missing tracking or guarantee that every AI-related visit will be identified.
How to find AI traffic in Google Analytics
Start with the Traffic acquisition report:
- Choose your website’s GA4 property. Confirm you are viewing the right business and site.
- Open Reports, then Traffic acquisition. It commonly appears under Acquisition. Customized report menus can look different.
- Select a recent date range and Session default channel group. Find the AI Assistant row to see recognized AI traffic for that period.
- Break down the sources. Apply a report filter that includes Session default channel group exactly matching AI Assistant, then use Session source / medium to see the individual sources.
If Traffic acquisition is missing from the menu, someone with Editor access or above can add it back. If the AI Assistant row is absent, check the dates, tracking, and report settings before assuming the feature is unavailable.
An editor can then save a dedicated report so you can return to it without rebuilding the filters each time.
Build an AI traffic report around useful questions
A source total is a starting point. For an owner, the next question is whether those visits help you understand interest in the business.
We recommend organizing the information around these decisions:
| Business question | What to review | How it helps |
|---|---|---|
| Which AI sources send visitors? | Sessions by source, using the AI Assistant filter | Shows where recognizable referrals originate |
| Which content brings them in? | Landing pages, meaning the first page in a visit | Identifies pages worth examining for relevance and a clear next step |
| Do visits lead to inquiries? | Verified inquiry events alongside traffic | Helps connect visits with actions that matter to the business |
For example, suppose a service page starts receiving ChatGPT referrals. You could review whether that page answers a buyer’s practical questions and makes it easy to contact you. That is a reason to inspect the page, not proof that every visitor is ready to buy.
At ClickIt CMO, we created AI Traffic by Source and AI Traffic Landing Pages. Both use the same native AI Assistant filter. We included sessions and engagement measures so we could review the sources and the pages visitors reached.
For that first setup, we focused on traffic and engagement. Before adding lead reporting, we would verify which events represent completed inquiries. Opening a contact page, clicking a button, and successfully submitting an inquiry are different actions. Check any event you report as a lead against what actually happened.
What the AI Assistant channel cannot tell you
Google AI Overviews and AI Mode are counted separately
Google’s default channel definitions place non-ad traffic from AI Overviews and AI Mode in Organic Search. Those Google Search experiences are excluded from the AI Assistant channel.
That means an AI Assistant report covers a specific category of visits. It is not a total for every way AI might influence someone to find your business.
Some visits arrive without usable source information
Analytics can only classify the information available to it. Redirects and other tracking issues can lose source details. Google’s explanation of Direct traffic describes why a visit may lack a clear referral source.
Some AI-influenced visits may therefore be difficult to identify. You cannot safely label an increase in Direct traffic as an increase in AI traffic.
Visits do not reveal every mention or recommendation
A person can read about your company in an AI answer without clicking through. A website traffic report cannot show that entire off-site journey.
Our AI Search Optimization services consider referral visits alongside other evidence, such as answer visibility, useful content, and qualified inquiries. Those measures answer different questions about how people discover and evaluate a business.
Keep your team’s website work out of the reports
Someone clicking a link in an AI answer and an AI assistant checking your website are different kinds of activity. The new channel does not automatically identify which visits belong to your team, developer, or assistant doing maintenance.
If you want to exclude that work, the controls need to fit your website and how people access it. GA4 supports internal traffic filtering based on IP addresses. Another approach is marking the browser used for website work and applying an appropriate collection control. Google documents ways to turn off Analytics collection when the required conditions are met.
We used a marked-browser approach for ClickIt’s own site work. It depends on marking that browser and maintaining the control when browsers or cookies change. It does not detect every automated visitor or clean up earlier visits already collected.
The setup should include two checks: confirm the intended work activity is excluded, and confirm ordinary visitor activity can still be recorded under the site’s consent settings. GA4’s Testing mode is useful when evaluating a data filter because activating an exclusion permanently removes matching incoming data from processing.
Need help setting up AI traffic reporting?
If you want these reports but do not know your way around GA4, we are happy to help. You also may know exactly what you want to measure and simply need someone to handle the setup.
ClickIt CMO can review your existing Analytics setup, create useful AI source and landing page reports, configure suitable work-traffic controls, and explain how to read the results. We will also identify whether inquiry tracking needs additional work before it can support lead reporting.
This can be a one-time project for an existing client or a business working with us for the first time. If you already have a marketing team, developer, or agency, we can coordinate with them and define who handles each part.
Ask ClickIt CMO for help tracking AI traffic in GA4. Tell us your website and what you would like the reports to help you understand. We will review the setup with you and agree on the work needed.
Frequently Asked Questions About How to Track AI Traffic in GA4
Will GA4 show the exact question someone asked ChatGPT?
Standard acquisition reports show the source of a visit and activity recorded on your website. They don’t show the visitor’s private conversation with an AI assistant. Use the landing page as context for what interested someone, without assuming you know their exact question.
How often should I review AI traffic reports?
A monthly review is a practical starting point for an owner, with closer checks after tracking changes. When there are only a few recognized AI visits, use a longer reporting period before comparing platforms or drawing conclusions from percentages.
Do I need a separate plugin to track AI traffic in GA4?
No separate plugin is required for GA4’s built-in AI Assistant channel. If your website already has working GA4 tracking, recognized AI referrals can be grouped automatically. You may still need help creating useful reports, checking inquiry tracking, and excluding your team’s website work from the data.
Why does my GA4 report show little or no AI traffic?
Your website may have few recognized AI referrals during the selected period, or visits may arrive without usable source information. Missing referral details can cause traffic to appear under Direct. Check your tracking, date range, and report filters before drawing conclusions. An empty report alone cannot establish that AI tools never mention your business.
Can I see whether AI traffic generates leads?
Yes, when your inquiry actions are correctly tracked. GA4 lets you mark important actions as key events and evaluate them across traffic channels. For a service business, that might be a successfully submitted inquiry form. Test the event first so a button click or an abandoned form is not counted as a completed inquiry.
Can ClickIt CMO set up AI traffic tracking as a one-time project?
Yes. We can review your existing GA4 setup, create AI traffic reports, configure appropriate work-traffic exclusions, and explain how to use the results. You can work with us on this as a one-time project, whether you are an existing client or contacting us for the first time. We can also coordinate with your current developer or marketing agency.
Need Help Setting Up AI Traffic Reporting?
Tell us your website and what you would like the reports to help you understand.
